About

Jennifer Kao is an assistant professor at the UCLA Anderson School of Management, where she co-teaches business strategy to MBA students. Her research investigates the relationship between firm strategies and innovation outcomes, particularly within the healthcare sector. She employs quantitative methods to examine how research organizations make strategic decisions related to acquiring, integrating, and disclosing information. Additionally, her work explores how organizations adapt their R&D and market entry decisions in response to regulation. Kao holds a Ph.D. in public policy from Harvard University, where she was a National Bureau of Economic Research Predoctoral Fellow in the Economics of Health and Aging and the International Network on the Value of Medical Research. She earned a B.S. in business administration from UC Berkeley and an M.Sc. in economics from the London School of Economics. Her research has been funded by notable organizations including the Society of Hellman Fellows and the National Institute of Health Care Management Foundation.

Research topics

• Computer Science
• Artificial Intelligence
• Psychology
• Neuroscience
• Machine Learning
• Algorithm
• Speech recognition
• Biology
• Cognitive psychology
• Medicine
• Anatomy

Selected publications

• Re-evaluating Position and Velocity Decoding for Hand Pose Estimation with Surface Electromyography

  ArXiv.org · 2026-03-09

  articleOpen accessSenior author

  Recent progress in real-time hand pose estimation from surface electromyography (sEMG) has been driven by the emg2pose benchmark, whose original baseline study concluded that velocity decoding outperforms position decoding in both reconstruction accuracy and trajectory smoothness. We revisit that conclusion under the original causal evaluation protocol. Using the same core architecture but a more stable training recipe, we show that position decoding models were previously underestimated because they are highly sensitive to a previously unswept decoder output scalar and can otherwise collapse into low movement solutions. Once this scalar is tuned, position decoding outperforms velocity decoding on the Tracking task across all three emg2pose generalization conditions, consistent with greater robustness to error accumulation. On the Regression task, the gap between position and velocity decoding is much smaller; instead, the largest gains come from multi-task training with Tracking, suggesting that the Regression objective alone does not sufficiently constrain the learned dynamics. Although position decoding models exhibit greater local jitter, a causal speed-adaptive filter preserves their accuracy advantage while yielding a more favorable smoothness-accuracy tradeoff than velocity decoding. Altogether, our results revise the original emg2pose modeling conclusions and establish a new state of the art among published streaming-compatible models on this benchmark.

  PublisherOA PDF

• DiSCo: Diffusion Sequence Copilots for Shared Autonomy

  2026-03-10

  articleOpen accessSenior author

  Shared autonomy combines human user and AI copilot actions to control complex systems such as robotic arms. When a task is challenging, requires high dimensional control, or is subject to corruption, shared autonomy can significantly increase task performance by using a trained copilot to effectively correct user actions in a manner consistent with the user’s goals. To significantly improve the performance of shared autonomy, we introduce Diffusion Sequence Copilots (DiSCo): a method of shared autonomy with diffusion policy that plans action sequences consistent with past user actions. DiSCo seeds and inpaints the diffusion process with user-provided actions with hyperparameters to balance conformity to expert actions, alignment with user intent, and perceived responsiveness. We demonstrate that DiSCo substantially improves task performance in simulated driving and robotic arm tasks. Project website: https://sites.google.com/view/disco-shared-autonomy/

  PublisherOA PDFDOI

• Spurious alignment between large language models and brains can emerge from non-robust methods and overlooked confounds

  Nature Communications · 2026-04-27

  articleOpen accessSenior author

  Emerging research seeks to draw neuroscientific insights from the neural predictivity of large language models (LLMs). However, as results rapidly proliferate, there is a growing need for large-scale assessments of their robustness. Here, we analyze a wide range of models and methodological approaches across three widely used neural datasets. We find that the use of shuffled train-test splits has contributed to findings that are influential but spurious. Furthermore, how activations are extracted from LLMs can bias results in favor of specific model classes. Lastly, we find that confounding variables, particularly positional signals and word rate, perform competitively with trained LLMs and fully account for the neural predictivity of untrained LLMs on these neural datasets. Although many studies in the field avoid these pitfalls, our results indicate that some apparent alignment between LLMs and brains has emerged from non-robust methods and overlooked confounds.

  PublisherOA PDFDOI

• DiSCo: Diffusion Sequence Copilots for Shared Autonomy

  arXiv (Cornell University) · 2026-03-24

  preprintOpen accessSenior author

  Shared autonomy combines human user and AI copilot actions to control complex systems such as robotic arms. When a task is challenging, requires high dimensional control, or is subject to corruption, shared autonomy can significantly increase task performance by using a trained copilot to effectively correct user actions in a manner consistent with the user's goals. To significantly improve the performance of shared autonomy, we introduce Diffusion Sequence Copilots (DiSCo): a method of shared autonomy with diffusion policy that plans action sequences consistent with past user actions. DiSCo seeds and inpaints the diffusion process with user-provided actions with hyperparameters to balance conformity to expert actions, alignment with user intent, and perceived responsiveness. We demonstrate that DiSCo substantially improves task performance in simulated driving and robotic arm tasks. Project website: https://sites.google.com/view/disco-shared-autonomy/

  PublisherDOI

• LightBeam: An Accurate and Memory-Efficient CTC Decoder for Speech Neuroprostheses

  arXiv (Cornell University) · 2026-03-14

  preprintOpen accessSenior author

  A promising pathway for restoring communication in patients with dysarthria and anarthria is speech neuroprostheses, which directly decode speech from cortical neural activity. Two benchmarks, Brain-to-Text '24 and '25, released intracranial recordings from patients with dysarthria along with a baseline algorithm trained with Connectionist Temporal Classification (CTC). Despite significant innovation on these benchmarks, all leading published prior work relies on a WFST-based CTC decoder that requires ${\sim}$320 GB of RAM. These memory requirements limit accessibility for both patients and researchers. Here, we propose LightBeam, a non-WFST based CTC decoder that requires only ${\sim}$10 GB of RAM and achieves state-of-the-art performance on both benchmarks. LightBeam achieves this by integrating an LLM into the beam-search process via delayed fusion, obviating the prior need for using a large N-gram LM. LightBeam is implemented in Python and is open-source.

  PublisherDOI

• LightBeam: An Accurate and Memory-Efficient CTC Decoder for Speech Neuroprostheses

  ArXiv.org · 2026-03-14

  articleOpen accessSenior author

  A promising pathway for restoring communication in patients with dysarthria and anarthria is speech neuroprostheses, which directly decode speech from cortical neural activity. Two benchmarks, Brain-to-Text '24 and '25, released intracranial recordings from patients with dysarthria along with a baseline algorithm trained with Connectionist Temporal Classification (CTC). Despite significant innovation on these benchmarks, all leading published prior work relies on a WFST-based CTC decoder that requires ${\sim}$320 GB of RAM. These memory requirements limit accessibility for both patients and researchers. Here, we propose LightBeam, a non-WFST based CTC decoder that requires only ${\sim}$10 GB of RAM and achieves state-of-the-art performance on both benchmarks. LightBeam achieves this by integrating an LLM into the beam-search process via delayed fusion, obviating the prior need for using a large N-gram LM. LightBeam is implemented in Python and is open-source.

  PublisherOA PDF

• Re-evaluating Position and Velocity Decoding for Hand Pose Estimation with Surface Electromyography

  Open MIND · 2026-03-09

  preprintSenior author

  Recent progress in real-time hand pose estimation from surface electromyography (sEMG) has been driven by the emg2pose benchmark, whose original baseline study concluded that velocity decoding outperforms position decoding in both reconstruction accuracy and trajectory smoothness. We revisit that conclusion under the original causal evaluation protocol. Using the same core architecture but a more stable training recipe, we show that position decoding models were previously underestimated because they are highly sensitive to a previously unswept decoder output scalar and can otherwise collapse into low movement solutions. Once this scalar is tuned, position decoding outperforms velocity decoding on the Tracking task across all three emg2pose generalization conditions, consistent with greater robustness to error accumulation. On the Regression task, the gap between position and velocity decoding is much smaller; instead, the largest gains come from multi-task training with Tracking, suggesting that the Regression objective alone does not sufficiently constrain the learned dynamics. Although position decoding models exhibit greater local jitter, a causal speed-adaptive filter preserves their accuracy advantage while yielding a more favorable smoothness-accuracy tradeoff than velocity decoding. Altogether, our results revise the original emg2pose modeling conclusions and establish a new state of the art among published streaming-compatible models on this benchmark.

  DOI

• Complementary cortical and striatal encoding of locomotor preparation and performance

  iScience · 2025-11-20 · 1 citations

  articleOpen access

  Cortical and striatal circuits encode information related to both motor initiation and execution. While neural dynamics in these areas show substantial similarities, possibly reflecting shared information content, studies directly comparing the cortical and striatal encoding of locomotor preparation and performance have been lacking. Here, we contrasted the neural coding properties of mouse primary motor and medial prefrontal cortex, as well as dorsolateral and dorsomedial striatum, prior to and during bouts of self-paced locomotion in an open field. All four areas contained cells active during both the preparatory and performance periods of locomotion. However, the decoding of behaviorally relevant information using population-level activity revealed significant regional variations. Specifically, dorsomedial striatum more accurately encoded the preparatory period prior to walking, while primary motor cortex, followed by dorsolateral striatum, more accurately encoded rhythmic limb kinematics during walking. Together, this work provides evidence for a complementary neural coding scheme for locomotor preparation and performance in cortical and striatal circuits.

  PublisherDOI

• Multiregional representations of intertemporal decision making in human single neurons

  Scientific Reports · 2025-07-08

  articleOpen access

  Understanding the neural mechanisms underlying delay discounting-the tendency to prefer smaller, immediate rewards over larger, delayed rewards-is critical for elucidating the etiology of impulsive decision-making, a hallmark of several psychiatric conditions including substance use and impulse control disorders. Here, we investigate single-neuron activity in the orbitofrontal cortex (OFC), hippocampus, and amygdala of nine human participants performing a delay discounting task. Intracranial recordings yielded a total of 193 single units (50 OFC, 68 amygdala, and 75 hippocampus) and reveal distinct neural correlates of decision-making, including representations of choice preferences and decision difficulty across all three regions. Analyses demonstrate preferential encoding of choice in the OFC. Additionally, we report that hippocampal activity reflects interindividual differences in discounting rates, with stronger representation observed in participants with slower temporal discounting. These findings provide novel insights into the multiregional neural computations underlying intertemporal decision-making and their relationship to impulsive behaviors.

  PublisherOA PDFDOI

• Neural basis of cooperative behavior in biological and artificial intelligence systems

  Science · 2025-09-25 · 5 citations

  articleCorresponding

  Cooperation, the process through which individuals work together to achieve common goals, is fundamental to human and animal societies and increasingly critical in artificial intelligence (AI). In this study, we investigated cooperation in mice and AI systems, examining how they learn to actively coordinate their actions to obtain shared rewards. We identified key social behavioral strategies and decision-making processes in mice that facilitate successful cooperation. These processes are represented in the anterior cingulate cortex (ACC), and ACC activity causally contributes to cooperative behavior. We extended our findings to AI systems by training artificial agents in a similar cooperation task. The agents developed behavioral strategies and neural representations reminiscent of those observed in the biological brain, revealing parallels between cooperative behavior in biological and artificial systems.

  PublisherDOI

Recent grants

• Next generation brain-machine interfaces controlled synergistically with artificial intelligence

  NIH · $2.3M · 2020–2025

• CAREER: Elucidating principles of cortical computation with recurrent neural networks

  NSF · $575k · 2020–2026

Frequent coauthors

• Krishna V. Shenoy

  Howard Hughes Medical Institute

  72 shared

• Stephen I. Ryu

  Korea Research Institute of Bioscience and Biotechnology

  33 shared

• Sergey D. Stavisky

  University of California, Davis

  27 shared

• Chethan Pandarinath

  The Wallace H. Coulter Department of Biomedical Engineering

  25 shared

• Paul Nuyujukian

  Stanford University

  23 shared

• Stephen I. Ryu

  Stanford University

  23 shared

• Leigh R. Hochberg

  Harvard University

  20 shared

• Chandramouli Chandrasekaran

  19 shared

Awards & honors

• Society of Hellman Fellows
• National Institute of Health Care Management Foundation